OpenAI Adds Interactive Charts, Forms and Buttons to ChatGPT With 'Intelligent UI'

OpenAI is rolling out a feature called Intelligent UI that lets ChatGPT respond to questions with interactive visuals rather than text alone. The update gives the chatbot the ability to pair a written answer with diagrams, charts, fillable forms and tappable buttons, according to the company's blog post announcing the change. It is available to all users and ships alongside GPT-6, the next version of OpenAI's flagship model.

The shift matters because ChatGPT's output has been essentially flat text since it launched, with images and tables appearing as static blocks inside the chat window. Intelligent UI turns those blocks into components a user can interact with — pressing a button, entering data in a form, or adjusting a chart — without leaving the conversation. That moves the product closer to a lightweight app surface than a chat box, a direction rivals have also been probing.

OpenAI did not detail which chart types, form fields or button actions will be supported at launch, nor whether developers can define custom interactive elements through the API. The blog post frames the feature as part of the GPT-6 release rather than a standalone product, which suggests OpenAI sees generated interfaces as core model capability rather than a bolt-on. The company has not published a full rollout timeline beyond saying availability is expanding to all users.

For builders, the immediate question is whether Intelligent UI will be exposed programmatically. If ChatGPT can generate a working form or chart from a prompt, the same capability could eventually replace some hand-built front ends for internal tools, dashboards and data-entry workflows. That would put OpenAI in more direct competition with low-code platforms and business-intelligence vendors, not just other chatbot makers.

The update also raises unresolved issues around accuracy and control. A chart generated from a model's own reasoning can look authoritative while resting on numbers the model inferred rather than retrieved, and interactive forms invite users to submit data through an interface OpenAI has not described in detail. How the company labels machine-generated visuals, and whether it cites the underlying data, will shape how much enterprises trust the feature.

Watch for the API documentation and any developer tooling OpenAI releases next, since that will determine whether Intelligent UI stays a consumer convenience or becomes a building block for third-party products. Also worth tracking is how Google and Anthropic respond — both have shipped multimodal features, but neither has made generated interactive interfaces a headline capability. The competitive pressure now runs toward models that do more than talk back.